IVIvy Cerelia Valerie

Ivy Cerelia Valerie

*This is an algorithmically generated author record based on publicly available information.

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article

Previous Organizations / Institutes

The organizations / institutes from the scholar's other articles in the past ten years


Ivy Cerelia Valerie has a total of 27 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Bridging Clinical Observational Research and Data Standards in Head and Neck Cancer Care: Scoping Review and Mapping of Clinical Data Elements

    by Ivy Cerelia Valerie, Farhannah Aly, Ghazaleh Niknam Shirvan, Fahim Irfan Alam, Meredith Johnston, Benjamin Y. Kong, Merran Findlay, Lois Holloway, Georgina Kennedy

    Purpose: A scoping review and mapping of clinical data elements in observational studies were conducted to assess head and neck cancer (HNC) clinical data, alignment with existing observational oncology data standards, and the demonstrated opportunities for using these data standards in observational studies. Methods: The scoping review was performed to systematically find and extract data elements from real-world studies emphasizing HNC treatment patterns. Text-embedding–assisted and manual mapping was performed to map data elements to their representations in HNC Ontology, Operational Ontology for Oncology, Minimal Common Oncology Data Elements, and the Observational Medical Outcomes Partnership Common Data Model. Study characteristics, data elements and their mapping, and proportion of clinical domains represented were analyzed. Demonstrated opportunities for using standard data elements were reviewed based on the mapping results. Results: A total of 98 observational studies were included. National databases were the most common data source (59.2%), and disease trajectory outcomes were the most frequently reported study end point (95.9%). Overall, 1,613 data elements were reported and classified into five domains: diagnosis (1.2%), health system (5.7%), tumor (27.2%), person (32.7%), and treatment (33.2%). Mapping coverage varied across domains, with 85%-100% coverage in diagnosis, 83.6%-95.2% in treatment, 85.3%-97.8% in tumor, 50.7%-89% in person, and 43.8%-81.2% in health system domains. Demonstrated opportunities for applying data standardization include facilitating geographical links to personal data elements, streamlining value derivation of longitudinally tracked data elements, and deriving complex treatment constructs. Conclusion: The existing data standards and associated terminology systems covered most of the clinical observational data elements, but gaps remain in structuring complex treatment factors that may affect patient outcomes. Resulting mapping offers generalizable representations across standards, laying the groundwork for further refinement through collaborative data element curation from local conventions under the guidance of data standard experts.

  • Open Access

    A Schema for Extraction of Comprehensive

    by Ivy Cerelia Valerie, Patrick R. Alba, Georgina Kennedy

    The characterization of patient smoking history is an integral part of evidence-based cancer care. We developed a schema to fully describe longitudinal history and quantified behavioral patterns for smoking consumption, including substance, modality, recency, quantity and cessation efforts, to support extraction from free-text patient histories.

  • Open Access

    Property Graph for Exploring Similarities in Real-World Oncology Standards

    by Ivy Cerelia Valerie, Lois Holloway, Timothy Churches, Georgina Kennedy

    Real-world oncology standards have reached a high level of maturity, yet documentation about the relationship between standards is limited. Mapping standards similarities is the foundational step towards enabling data integration and sharing across different standards. Entity mapping is an iterative process to find corresponding elements in different semantic spaces with incremental specificity. It can identify areas of overlap and complement, thereby facilitating manual and automated integration efforts. We aim to provide a baseline description by comparing the entities and standard values in real-world oncology standards using property graph. Empirical exploration using property graph showed that all standards shared entities that represent important oncology domains spanning across demographic, diagnostic, and therapeutic details. Overlapping and complementary entities and values were reviewed via direct matching and automated calculation of text embedding similarity. Property graph effectively modeled complex oncology standards while enabling contextual understanding through relationship exploration. The considered approach demonstrated the feasibility of integrating diverse sources and interoperable sharing of artefacts.

  • Open Access

    Current Challenges in Digital Representation of Variation in Cancer Care

    by Sophia Kee, Ivy Cerelia Valerie, Georgina Kennedy, Merran Findlay, Timothy Churches, Alexandra Vassar

    Advances in cancer treatment have improved patient outcomes and survival in recent decades. Increased complexity, duration, and individualisation of treatment protocols present an important challenge for care teams monitoring adherence to best-practice care. A rigid rules-based system for flagging outliers is not fit for purpose, as there are sound reasons for deviating from baseline protocols, such as the management of treatment side effects to a tolerable degree, however the methods for determining the bounds of appropriateness for variation are not well studied or understood. The development of digital representations to inform cancer care delivery in a timely and continuing manner is crucial. This scoping review seeks to identify gaps in current methods and propose a novel approach to digitally represent patient journeys in clinically meaningful visual and computational forms. These methods can be combined to produce real-time, clinically applicable tools such as group-level business-intelligence dashboards (are processes and resources adequate to ensure that patients are being treated according to best practice?) as well as individual-level decision support (what is the likely outcome for this patient if treatment is stopped early based on prior data?) and day to day clinical workflows (what has happened to this patient so far?)

  • Open Access

    Enriching Data Science and Health Care Education: Application and Impact of Synthetic Data Sets Through the Health Gym Project

    by Nicholas I-Hsien Kuo, Oscar Perez-Concha, Mark Hanly, Emmanuel Mnatzaganian, Brandon Hao, Marcus Di Sipio, Guolin Yu, Jash Vanjara, Ivy Cerelia Valerie, Juliana de Oliveira Costa, et al.

    Large-scale medical data sets are vital for hands-on education in health data science but are often inaccessible due to privacy concerns. Addressing this gap, we developed the Health Gym project, a free and open-source platform designed to generate synthetic health data sets applicable to various areas of data science education, including machine learning, data visualization, and traditional statistical models. Initially, we generated 3 synthetic data sets for sepsis, acute hypotension, and antiretroviral therapy for HIV infection. This paper discusses the educational applications of Health Gym’s synthetic data sets. We illustrate this through their use in postgraduate health data science courses delivered by the University of New South Wales, Australia, and a Datathon event, involving academics, students, clinicians, and local health district professionals. We also include adaptable worked examples using our synthetic data sets, designed to enrich hands-on tutorial and workshop experiences. Although we highlight the potential of these data sets in advancing data science education and health care artificial intelligence, we also emphasize the need for continued research into the inherent limitations of synthetic data.


  • Open Access

    Enriching Data Science and Health Care Education: Application and Impact of Synthetic Data Sets Through the Health Gym Project (Preprint)

    by Nicholas I-Hsien Kuo, Oscar Perez-Concha, Mark Hanly, Emmanuel Mnatzaganian, Brandon Hao, Marcus Di Sipio, Guolin Yu, Jash Vanjara, Ivy Cerelia Valerie, Juliana de Oliveira Costa, et al.

    UNSTRUCTURED Large-scale medical data sets are vital for hands-on education in health data science but are often inaccessible due to privacy concerns. Addressing this gap, we developed the Health Gym project, a free and open-source platform designed to generate synthetic health data sets applicable to various areas of data science education, including machine learning, data visualization, and traditional statistical models. Initially, we generated 3 synthetic data sets for sepsis, acute hypotension, and antiretroviral therapy for HIV infection. This paper discusses the educational applications of Health Gym’s synthetic data sets. We illustrate this through their use in postgraduate health data science courses delivered by the University of New South Wales, Australia, and a Datathon event, involving academics, students, clinicians, and local health district professionals. We also include adaptable worked examples using our synthetic data sets, designed to enrich hands-on tutorial and workshop experiences. Although we highlight the potential of these data sets in advancing data science education and health care artificial intelligence, we also emphasize the need for continued research into the inherent limitations of synthetic data.


  • Open Access

    Ferritin in pediatric critical illness: a scoping review

    by Ivy Cerelia Valerie, Anak Agung Sagung Mirah Prabandari, Dyah Kanya Wati

    This scoping review aimed to elucidate and summarize the predictive role of serum ferritin in critical pediatric illness. The Preferred Reporting Items for Systematic reviews and Meta-Analyses methodology was employed to conduct a scoping review of 5 databases (MEDLINE, CENTRAL, ProQuest, ScienceDirect, and Epistemonikos) from the date of inception through January 24, 2022. Primary research studies involving subjects aged <18 years and serum ferritin levels were screened and reviewed independently following an a priori defined protocol. Of the 1,580 retrieved studies, 66 were analyzed. Summary statistics of serum ferritin levels for overall and condition-specific studies were reported in 30 (45.4%) and 47 studies (71.2%), respectively. The normal range was defined in 16 studies (24.2%), whereas the threshold was determined in 43 studies (65.1%). A value of <500 ng/mL was most often the upper limit of the normal range. Serum ferritin as a numerical variable (78.9%) was usually significantly higher (80.8%) in the predicted condition than in controls, while as a categorical variable with preset thresholds, ferritin was a significant predictor in 84.6% of studies. A total of 22 predictive thresholds predicted mortality (12 of 46 [26.1%]), morbidity (18 of 46 [39.1%]), and specific (16 of 46 [34.8%]) outcomes in 15 unique conditions. Increased precision in serum ferritin measures followed by close attention to the threshold modeling strategy and reporting can accelerate the translation from evidence to clinical practice.